Study on Molecular Mechanism of Liver Metastasis in Patients with Late Colorectal Cancer by NGS Technique

Jing WU, Bin HU, Juanna DONG, Xiuqiang LIN, Weiying DENG, Wei WANG

Foshan First People’s Hospital, Foshan 528010, China

Abstract [Objectives] To analyze the difference of gene expression between primary lesion and metastatic lesion in patients with advanced colorectal cancer with liver metastasis by NGS detection, and to further explore the molecular mechanism of liver metastasis in patients with advanced colorectal cancer. [Methods] In a case-control study combined with cohort study, NGS sequencing was performed in 9 patients with postoperative first-line cetuximab combined with chemotherapy, and retrospective analysis was made to find the biomarker that can predict the curative effect; then 9 samples of biomarker were prospectively verified; at the same time, 9 cases of primary lesions (2 sites) and 9 cases of liver metastases were detected by WES. [Results] The mutation type of primary lesion was mainly Missense (missense mutation): a total of 1 802 mutations were detected in the primary lesion samples of 9 patients, of which Missense accounted for 88.01%; a total of 2 150 mutations were detected in the metastatic samples of 9 patients, of which Missense accounted for 87.67%. The results of Venn diagram showed that there were 1 010 common mutations in primary and metastatic lesions, accounting for 42.4%. There were 819 genes in metastatic lesions, accounting for 34.4%; there were 553 unique genes in the primary lesion, accounting for 23.2%. The results of thermography showed that there were 7 mutant genes with more than 33.3% primary lesions, including APC, TTN, HDAC2, RELN, RHPN2, RYR3 and TP53; there were 13 mutant genes with metastatic lesions of more than 33.3%, including APC, RHPN2, TTN, TP53, COBLL1, DNAH12, DNAH17, EVC2, RYR3, SVEP1, SYNE1, ZNF197 and ZNF423, and the mutation rate of APC increased from 66.7% to 77.8%; TTN increased from 44.4% to 66.7%; RHPN2 rose from 33.3% to 66.7%; TP53 rose from 33.3% to 55.6%. COBLL1, DNAH17 and EVC2 rose from 22.2% to 33.3%. SVEP1, SYNE1, ZNF197 and ZNF423 were only expressed in metastatic lesions, and the expression increased from 0% to 33.3%. The results of GO enrichment showed that the top five mutant genes in colon cancer were KRAS, BRAF, TP53, APC and PI3KCA. [Conclusions] The primary and metastatic lesions were mainly Missense; APC, TP53 mutation was a common mutant gene with high expression; the expression of PHPN2 and TTN mutant genes increased by 33.3% and 11.1%, respectively; the mutation rate of BRAF gene was the second in GO enrichment.

Key words Colorectal cancer, NGS, Liver metastasis

1 Introduction

Colorectal cancer is the most common malignant tumor in China, with a high incidence in the top five malignant tumors, posing a serious threat to people’s health. The liver is the most common organ for colon cancer metastasis, about half of the patients have liver metastasis, and 10%-25% of colon cancer patients have synchronous liver metastasis when the diagnosis is established[2]. At present, surgical resection is still the only treatment for liver metastasis. Although progress has been made in chemotherapy and radiotherapy, most patients will relapse after hepatectomy[3]. Metastasis to liver is a series of extremely complex and highly selective processes, in which many related factors have been evaluated[4]. Hematogenous metastasis is one of the possible mechanisms of colorectal cancer with liver metastasis at present. In addition, tumor cells metastatic to the liver will undergo some genetic changes to complete the process of transformation, growth, angiogenesis, invasion, transmission and liver metastasis. Through the review of the literature, we found that there are 20 molecular factors related to proteolysis, cell adhesion, angiogenesis and cell survival related to colorectal cancer with liver metastasis. Although there are many studies on the molecular mechanism of colorectal cancer with liver metastasis, the mechanism of metastasis is still unclear. In-depth study of the pathological features and molecular mechanism of colorectal cancer with liver metastasis can provide a theoretical basis and a new target for individualized treatment for early detection of liver metastasis, prevention and treatment of colorectal cancer with liver metastasis.

Next generation sequencing (NGS) is a new sequencing technology emerged in recent years, which has a large sequencing flux and can simultaneously detect the DNA base sequences of specific samples of tumor patients. It is an efficient means to find known and unknown disease-related gene variations[8]. In this paper, NGS technique was used to detect the molecular map of primary and metastatic lesions in patients with advanced colon cancer with liver metastasis, to explore the molecular mechanism of liver metastasis in patients with advanced colon cancer by analyzing the difference of gene expression between primary and metastatic lesions, and to find molecular markers for predicting therapeutic efficacy and prognosis in accordance with curative effect and follow-up information of patients. It can provide a theoretical basis and a new target for individualized treatment for early detection of liver metastasis, prevention and treatment of colorectal cancer with liver metastasis.

2 Materials and methods

2.1 Criteria

2.1.1Inclusion criteria. The age is over 18 years, male or female; patients with advanced colorectal cancer with liver metastasis confirmed by pathology; the early pathological information of the patient is complete, and the tumor location and pathological type are complete; other indexes of the patients meet the general clinical trial conditions; NGS was used to detect colorectal cancer related genes, and there was no mutation of KRAS gene; first-line treatment with cetuximab combined with chemotherapy; a written informed consent signed by himself or his legal representative may be obtained.

2.1.2Exclusion criteria. Have a history of alcohol or drug abuse; female patients during pregnancy; any signs of severe or uncontrolled systemic disease that researchers believe may significantly change the patient’s risk/benefit balance; those who are considered unsuitable for inclusion by researchers.

2.2 Research methodsIn a case-control study combined with cohort study, NGS sequencing was performed in 9 patients with postoperative first-line cetuximab combined with chemotherapy, and retrospective analysis was made to find the biomarker that can predict the curative effect; then 9 samples of biomarker were prospectively verified; at the same time, 9 cases of primary lesions (2 sites) and 9 cases of liver metastases were detected by WES.

2.3 Research steps

2.3.1Patient enrollment and sample collection. A total of 9 patients with advanced colorectal cancer with liver metastasis were collected. In accordance with the requirements of hospital ethics, the attending doctor shall give informed consent to the patients and their families, fully discuss the advantages and disadvantages of participating in the project, and obtain the informed consent form signed by the patients and their families in writing. The primary lesion (2 loci) and metastatic lesion of tissue obtained by surgery/tissue obtained by needle biopsy were taken from 9 patients, and the paracancerous tissue and peripheral blood were taken as negative control. If it was fresh tumor tissue obtained by surgery, it should be put into the tissue preservation solution in the sample collection box; the paraffin sample was sliced with a thickness of 6-10 μm and stored in a slice box. It was first be confirmed by pathological testing that the tumor cell content should not be less than 20%, and the sample was transported to the central laboratory for preservation at room temperature. 10 mL of whole blood samples were collected by EDTA anticoagulant tube, and the plasma was separated by centrifugation (within 2 h). The precipitated cells were preserved and transported to the central laboratory and cryopreserved at -80 ℃. Aseptic centrifuge tube (50 mL) was used to collect 50 mL of ascites samples (no less than 25 mL) and transported to the central laboratory for preservation at room temperature.

2.3.2High-throughput sequencing. QIAGEN DNA extraction kit was used to extract DNA from tissue samples and precipitated cell samples, and QIAGEN plasma DNA extraction kit was used to extract ctDNA from plasma samples. After extraction, Qubit 3.0 was used for quantitative testing and quality control. The DNA of tissue and blood precipitated cells were segmented by ultrasound, the ends of fragmented DNA and ctDNA were repaired. A base was added, and Illumina sequencing connector was added to construct a high-throughput sequencing library. Geneseeq intestinal cancer driving gene panel was used to enrich the target region. After enrichment, the library was put on the Illumina Hiseq sequencing platform, and the Illumina PE150 sequencing kit was used for high-throughput sequencing. The average sequencing depth of blood precipitated cells was not less than 100×; the average sequencing depth of tissue samples was not less than 1 000×; the average sequencing depth of ctDNA and ascites samples was not less than 5 000×.

2.3.3Bioinformatics analysis. The gene fusion mutation, gene copy number change, point mutation, base insertion/deletion,etc.of tumor tissue DNA, plasma and ascites ctDNA were comprehensively analyzed by bioinformatics, and the germline mutation was removed by comparing with the negative control of blood cells, and the tumor-specific mutation was screened out.

3 Results and analysis

3.1 Mutation typeThe mutation type of primary lesion was mainly Missense (missense mutation): a total of 1 802 mutations were detected in the primary lesion samples of 9 patients, of which Missense accounted for 88.01%, Nonsense accounted for 7.99%, Frameshift accounted for 2.33%, Splicing accounted for 1.11%, and InFrame_indel accounted for 0.28% (Fig.1A).

The mutation type of metastatic lesion was mainly Missense (missense mutation): a total of 2 150 mutations were detected in the metastatic lesion samples of 9 patients, of which Missense accounted for 87.67%, Nonsense accounted for 8.09%, Frameshift accounted for 2.6%, Splicing accounted for 1.16%, and InFrame_indel accounted for 0.28% (Fig.1B).

Note: A. primary lesion; B. metastatic lesion.

3.2 Venn diagramThere were 1 010 common mutation genes (42.4%) in primary lesions and metastatic lesions after removing duplications, accounting for 42.4%; there were 819 genes in metastatic lesion, accounting for 34.4%; there were 553 genes unique to the primary lesion, accounting for 23.2% (Fig.2).

Fig.2 The proportion of gene mutations between primary and metastatic lesions

3.3 Heat mapFig.3 shows that there were 7 mutant genes with more than 33.3% of mutations in primary lesions, including APC, TTN, HDAC2, RELN, RHPN2, RYR3 and TP53; there were 13 mutant genes with metastasis over 33.3%, including APC, RHPN2, TTN, TP53, COBLL1, DNAH12, DNAH17, EVC2, RYR3, SVEP1, SYNE1, ZNF197, ZNF423, and the mutation rate of APC increased from 66.7% to 77.8%, an increase of 11.1%; TTN rose from 44.4% to 66.7%, an increase of 22.3%; RHPN2 rose from 33.3% to 66.7%, an increase of 33.4%; TP53 rose from 33.3% to 55.6%, an increase of 22.3%; COBLL1, DNAH17 and EVC2 rose from 22.2% to 33.3%, an increase of 11.1%; SVEP1, SYNE1, ZNF197 and ZNF423 were only expressed in metastatic lesions, and the expression increased from 0 to 33.3%.

Note: A. primary lesion; B. metastatic lesion.

3.4 GO enrichment resultsThe enrichment of mutant genes in the database showed that the top five mutated genes in colorectal cancer were KRAS, BRAF, TP53, APC and PI3KCA (Fig.4).

Fig.4 Mutation map of colorectal cancer in database

4 Discussion

In recent years, the incidence of rectal cancer is increasing, and for early rectal cancer patients, timely treatment can ensure that the 5-year survival rate is higher than 90%, but for colorectal cancer patients with liver metastasis, the 5-year survival rate is lower than 10%[9-10]. The pathogenesis of colorectal cancer is complicated[11], and liver metastasis usually occurs in the late stage. The pathogenesis of liver metastasis is related to a variety of gene pathways and genetic mutations. Therefore, among the many factors inducing colorectal cancer, genetic factors are the important causes of colorectal cancer.

NGS-based WES sequencing is to study mutant exons in the coding region. It uses gene capture technology to obtain DNA in the exon region of the whole genome, and then carries out high-throughput sequencing after enrichment. It has the characteristics of high accuracy, more comprehensive detection of mutant genes, less procedures and low detection cost. It is easier to obtain the relevant information of tumor mutant genes, and is helpful to finding new cancer markers[12].

For this reason, this project used the second generation gene to detect WES, comprehensively analyzed the mutant gene of colorectal cancer with liver metastasis, and clarified its molecular mechanism. The study found that most of the mutation types of primary and metastatic lesions were missense mutations, accounting for about 88%; the Venn diagram analysis of 9 patients after removing repetitive mutations showed that there were 1 110 genes, accounting for about 43% of the total mutations, indicating that there was a high consistency of gene mutation between the primary lesions of colorectal cancer and synchronous liver metastases.

Genetic mutations may lead to abnormal expression of some proteins and lead to hereditary colorectal cancer[13]. Through heat map analysis, it was found that more than 33.3% of the mutant genes in the primary lesions were APC, TTN, HDAC2, RELN, RHPN2, RYR3 and TP53. The genes that accounted for 33.3% in the metastatic lesions were APC, RHPN2, TTN, TP53, COBLL1, DNAH12 and DNAH17. The common mutant genes were APC, RHPN2, TTN and TP53. Through GO enrichment analysis, it was found that APC and TP53 genes were also found in the top 5 mutant genes KRAS, BRAF, TP53, APC and PIK3CA, which showed that these two genes had high mutation frequency and existed in both primary and metastatic lesions, and showed an upward trend, indicating that liver metastasis of colorectal cancer could be predicted when APC, TP53 mutations were detected in primary lesions. This finding provides a theoretical basis for clinical judgment of colorectal cancer with liver metastasis and advanced colorectal cancer.

From the mutation map of primary and metastatic lesions, we can see that the gene with the highest mutation frequency is APC, with a mutation frequency of about 70% (mainly Frameshift and Nonsense mutations), with 4 and 3 mutation patients respectively. Frameshift mutation refers to a change in the reading mode of the triplet code, resulting in a change in the amino acid sequence of proteins, and a completely different translation of proteins. Nonsense mutation[14]means that base substitution makes a codon directly change into stop codon UAA, UAG and UGA, so that the translation of peptide chain here is terminated, resulting in incomplete peptide chain and protein, inactivation of protein and reduction of protein yield encoded by gene. APC is a tumor suppressor gene. The expressed APC protein can induce apoptosis of cancer cells, regulate cell migration, proliferation, adhesion and stabilize chromosomes to maintain normal cell cycle. The mutation mainly occurs in the exon region, the APC gene includes 15 exons[15], and exon 15 accounts for 77% of the coding gene, about 6 571 bp. The mutation at the 5’ end of exon 15 between codon 1 286 and codon 1 513 is closely related to sporadic colorectal cancer lesions. Another common mutation type of APC is frameshift mutation, which is caused by the deletion or insertion of small fragments of genes, mostly small fragments of 2 and 5 bp, and point mutations mostly occur between exons 5, 6, 8, 9 and 12-15[16]. These mutations lead to a completely different peptide chain, which makes the expressed APC protein abnormal and unable to interact with β-catenin and E-cadherin to regulate cell adhesion. At the same time, the mutant APC protein cannot interact with tubulin, affecting cell migration[17]. Therefore, APC mutation leads to cell adhesion-migration dysfunction[18], which is one of the causes of colorectal cancer with liver metastasis.

The TP53 gene exists in both the primary lesion and the metastatic lesion, and the mutation frequency of the primary lesion is 33.3%, while the mutation frequency of the metastatic lesion is as high as 55.6%, an increase of 22.3%. It can be seen that its mutation has a great impact on the occurrence of liver metastasis and is one of the main mutant genes leading to liver metastasis. TP53 is a tumor suppressor gene[19], which is located in chromosome 17P13.1 and has 11 exons. All of the mutation types are missense mutations. Missense mutation generally means that the codon base pair encoding a certain amino acid is replaced by other bases, changing from one codon to another codon, resulting in changes in the amino acid sequence of the translated polypeptide chain and the variation of the synthesized protein. Sometimes missense mutations have no effect on peptide sequences, but in most cases they are harmful and lethal. The missense mutation of TP53 occurs in the DNA binding region (amino acid 100-293), and the variation was mainly in exons 5, 6, 7 and 8. The mutation of exon 7 occurs at the third base of codon 249, which is mainly the SNV mutation of G>T, C>A. The C>T, G>A substitution mutation occurs at codon 1 in exon 8. TP53 is a negative regulatory factor in the cell growth cycle, which participates in the regulation of all aspects, and is responsible for the repair of DNA, cell growth and differentiation and other functions. The TP53 protein encoded by TP53 contains N-terminal transactivation domain, DNA binding core domain, tetramerization domain and C-terminal regulatory domain[20]. It can regulate various kinds of cell stress, cell growth, apoptosis and senescence. Its mutation or deletion will lead to genomic instability and excessive cell proliferation, induce tumorigenesis, and is one of the important factors that initiate the process of carcinogenesis.

In addition, the mutation rate of RHPN2 and TTN is also very high, 66.7% in the metastatic lesions. The former is Missense, the latter (1/3) is Nonsense, and the latter (2/3) is Missense. RHPN2 is an intracellular effector of RhoA protein pathway, which is not SMADs-dependent in TGF-β signaling pathway[21], regulating the expression of TGF-β target genes. TGF-β signaling pathway plays an important role in the genesis and metastasis of colorectal cancer. The expression of RHPN2 in normal human is 105, while the expression in cancer tissue is as high as 1 606. RHPN2 gene translates a Rho binding protein commonly expressed in human body. RHPN2 mutations, especially Rho binding site mutations, can block the binding with RhoA (GDP- binding with GTP-), promote mesenchymal transformation of rectal cancer cells, accelerate tumor cell metastasis, invasion and liver metastasis[22]. TTN is a gene that encodes myonectin[23-25]. Its mutation can lead to truncation of expressed myonectin, which is mainly caused by idiopathic or familial dilated cardiomyopathy. High frequency TTN mutations were detected in both primary and metastatic lesions of colorectal cancer. The main types of mutations were missense mutations and nonsense mutations, and the mutation frequency of metastatic lesions was significantly higher than that of primary lesions, which provided a new theoretical target for the detection of liver metastasis. The mutation rate of BRAF gene in GO enrichment was the second, accounting for about 12%, and all of them were missense mutations. BRAF is a proto-oncogene[26], which exists in the normal cell genome of organisms. It is an important signal molecule that regulates the downstream EGFR pathway and regulates cell growth, differentiation and apoptosis. BRAF mutation can activate the signal pathway composed of downstream RAS-RAF-MEK-ERK protein kinase of EGFR, affect the MITF factor of cell cycle, enhance the mitotic ability of cells, lead to abnormal cell proliferation and differentiation, and induce cancer cell formation. Among the many types of mutations, the most common mutation is V600E: a T>A mutation occurs at base 1 799 in exon 15 of BRAF, which turns the encoded valine into glutamate. After activation of kinase, the activity of mutant BRAF was 500 times higher than that of wild type BRAF. There is also a low frequency missense mutation, which is located on glycine in exon 11. BRAF is a more effective drug target than KRAS. Routine detection of BRAF gene mutation in colorectal cancer patients before treatment is of great significance for the use of EGFR targeted drugs and the prognosis of patients[27].

To sum up, through WES detection based on NGS, the gene mutation maps and GO enrichment analysis maps of primary and metastatic lesions were compared, and the clonal evolution of tumors was analyzed. It was found that the mutation rates of APC, RHPN2, TTN and TP53 in primary and metastatic lesions were the highest, which were the main causes of colorectal cancer with liver metastasis and disease-driven genetic mutation factors, and were important areas for gene detection of colorectal cancer with liver metastasis. In addition, TTN mutations have always been common in cardiomyopathy. High frequency TTN mutations were detected in metastatic lesions, which provided a new target for the detection of colorectal cancer with liver metastasis and targeted drug research. At the same time, the BRAF gene mutation rate in GO enrichment map was extremely high, which was a more effective target for detecting liver metastasis by colorectal cancer than KRAS. For its target, subsequent research and development of targeted drugs will provide patients with more choices on the basis of EGFR inhibitors in the treatment of rectal cancer. The sample of this study is limited, and NGS may use more comprehensive sequencing to identify the cause and find a more targeted treatment. Now NGS analysis is used to study the mechanism of gene mutation in a variety of cancers. This will help us to improve our understanding of mutations in cancer-causing genes, and discover more biomarkers and potential drug targets through NGS.